Aiming at the problem that the traditional robot can not grasp multiple types of workpieces by using fixed point teaching method, this paper designs an industrial robot grasping system based on intelligent vision and PLC. The overall scheme of the system is layered according to the IEC62264-1 standard. The visual module and cooperative robot are used as field equipment, PLC is the main controller of the system, HMI and SCADA are used for process monitoring, machine learning technology and a variety of algorithms are adopted, which can quickly and accurately identify the type and color of target workpiece, and realize the multi type workpiece capture of industrial robot and digital control of production unit. The results show that the system has good grasping accuracy, can effectively complete the grasping task of a variety of complex workpieces, effectively improve the production efficiency and quality, and provide strong support for the intelligent upgrading of industrial production.